# Fileset

[Apl173108.pdf](https://mdr.nims.go.jp/filesets/939dbfee-d534-4f52-b0c5-9437d9a74d6f/download)

## Creator

Konstantin Iakoubovskii, Nobutsugu Minami, Yeji Kim, Kanae Miyashita, Said Kazaoui, Balakrishnan Nalini

## Rights



## Other metadata

[Midgap luminescence centers in single-wall carbon nanotubes created by ultraviolet illumination](https://mdr.nims.go.jp/datasets/54cfabe0-2f5c-4b8a-b3f5-f9b8012f2699)

## Fulltext

Midgap luminescence centers in single-wall carbon nanotubescreated by ultraviolet illuminationKonstantin Iakoubovskii, Nobutsugu Minami,a� Yeji Kim, Kanae Miyashita,Said Kazaoui, and Balakrishnan NaliniNanotechnology Research Institute, National Institute of Advanced Industrial Science and Technology(AIST), 1-1-1 Higashi, Tsukuba, Ibaraki 305-8565, Japan�Received 23 May 2006; accepted 4 September 2006; published online 23 October 2006�The authors report the effect of ultraviolet �UV� illumination on optical properties of single-wallcarbon nanotubes �SWCNTs� isolated using various dispersants. It is demonstrated that even weakUV light ��1 mW/cm2� can irreversibly alter the SWCNT structure, thus resulting in theemergence of hitherto unknown, redshifted photoluminescence �PL� peaks with concomitantreduction in some of the original PL peaks. These UV-induced changes are characterized in detailand attributed to the creation of midgap PL centers. © 2006 American Institute of Physics.�DOI: 10.1063/1.2364157�Optoelectronic applications of single-wall carbon nano-tubes �SWCNTs� critically hinge on the development ofmeans to design and control relevant properties. One of suchimportant properties is photoluminescence �PL�, which is avery sensitive and versatile probe of electronic states of ma-terials. The analysis and tunability of PL from SWCNTs�Ref. 1–5� constitute an essential step leading towards theapplications of SWCNTs.The present letter reports an unprecedented method toshift PL peaks of SWCNTs by a very simple treatment,namely, shining weak ultraviolet �UV� light on dispersant-aided dispersion of SWCNTs. We show that this shift is trig-gered by the interaction of SWCNTs with UV-excited dis-persants which eventually creates one-dimensionally �1D�confined defect states.SWCNTs �CoMoCAT, purchased from Southwest Nano-technologies�, showing a narrow PL spectrum with fewerpeaks, have been chosen as a starting material to facilitatespectral analysis. Note that essentially the same results wereobtained for other SWCNTs such as HiPCO �Carbon Nano-technology Inc.�. Sample preparation and measurementswere performed at room temperature. Dispersants were dis-solved in D2O to increase IR transparency at a concentrationof 10 g/ l. As dispersants, sodium carboxymethylcellulose6�CMC� and sodium dodecylbenzenesulfonate �SDBS� weremainly used, while several others were also tried. ThenSWCNTs were added at a concentration of 0.3 g/ l, sonicatedfor 20 min �20 kHz, �130 W�, and ultracentrifuged for 5 hat �150 000 g. Films were prepared by casting. Some filmswere detached from the silica substrates and mechanicallystretched tenfold to partially align the SWCNTs.6,7PL spectra were recorded with a custom-made setup,equipped with a diode laser �660 nm, intensity �1 W/cm2�and a nitrogen-cooled InGaAs photodiode array. For polar-ization measurements, the laser was depolarized, and the po-larization of PL was analyzed in terms of the “alignmentratio” of PL signals polarized parallel or perpendicular to thestretch direction.6,7 PL excitation �PLE� maps were obtainedusing a Fluorolog FL3-2TRIAX spectrofluorometer.Sample illumination due to PL or absorption measure-ments did not induce any detectable photochromic changes.Changes were generated by in situ UV illumination with a20 W D2 lamp ��1 mW/cm2 at the sample surface�. Thelamp spectrum is mostly concentrated in the 190–400 nmrange peaking at 200 nm �see inset in Fig. 3�. Film sampleswere kept in vacuum ��10−3 Torr� to avoid polymer photo-oxidation. The spot size and sample thickness were chosenso as to produce homogeneous illumination and induce neg-ligible sample heating.Most representative results were obtained using CMCdispersant, which produces concentrated, stable SWCNTdispersions.6 All results appeared very similar for solutionsand films and will mostly be presented for the film samplesunless otherwise mentioned.Figure 1 outlines a summary of the PL changes inducedin a CMC/CoMoCAT film by D2 lamp: illumination reducesPL intensity by approximately ninety percent in a time scale��15 min followed by saturation of changes. While PL in-tensity decreased, no significant spectral changes could bedetected after such short-term processing. Besides, interrupt-ing illumination at this stage resulted in full recovery ofPL �and absorption, vide infra� within 1 h. However, longera�Electronic mail: n.minami@aist.go.jpFIG. 1. �Color online� Summary of variations in PL spectrum �660 nmexcitation� from a CMC/CoMoCAT film induced by 20 W D2 lamp illumi-nation. No extra peaks were detected in the 1450–1640 nm range.APPLIED PHYSICS LETTERS 89, 173108 �2006�0003-6951/2006/89�17�/173108/3/$23.00 © 2006 American Institute of Physics89, 173108-1Downloaded 24 Oct 2006 to 150.29.237.184. Redistribution subject to AIP license or copyright, see http://apl.aip.org/apl/copyright.jsphttp://dx.doi.org/10.1063/1.2364157http://dx.doi.org/10.1063/1.2364157http://dx.doi.org/10.1063/1.2364157��2 h� illumination induced significant irreversible alter-ations, as shown by the “final” spectrum in Fig. 1: the peaksat 980, 1050, and 1150 nm weaken, and an additional strongsignal appears at �1235 nm. Note that these changes couldonly be observed a few hours after turning off D2 lamp, butnot during or immediately after illumination; they werestable for at least 6 months and could not be reversed byannealing �100–250 °C, 1 h�.More detailed and important information on the de-scribed changes was obtained by measuring PLE maps be-fore �Fig. 2�a�� and after �Fig. 2�b�� UV illumination. Thesemaps are generally interpreted as follows: PL peaks �X scale�correspond to the S11 transitions between the first pairs of thevan Hove singularities; the PLE peaks �Y scale� in the range800–1000 nm have been assigned to the phonon-assisted S11transitions,8 while those in the ranges 450–770 and270–400 nm are attributed to the S22 and S33 transitions,respectively. The S22/S11 pairs identify9 the SWCNT chiral-ity numbers as indicated in parentheses in Fig. 2�a�. Here it isimportant to note that all the PL peaks in the unilluminatedsample �Fig. 2�a�� exhibit different PLE spectra.The PLE map of Fig. 2�a� is drastically altered by D2lamp illumination: for the 573 nm excitation of the �6, 5�tube, Fig. 2�b� reveals that two additional PL peaks, labeledas X1 and X2, appear at 1125 and 1230 nm. These two peaksexhibit similar PLE spectra that also resemble that of theoriginal 1000 nm PL peak. A similar change occurs for the650 nm excitation of the �7, 5� tube. Two peaks emerge atthe longer wavelengths; however, the original 1050 nm peakconsiderably decreases here, apparently indicating PL inten-sity transfer from the latter to the former peaks. Remarkably,all the PLE features have comparable linewidth, but thewidth of PL lines is larger for the UV-induced than for the“original” PL peaks, thus resulting in the oval-shaped fea-tures in Fig. 2�b�.PL spectra similar to those in Fig. 1 �mainly due to �7, 5�tubes� were also recorded from stretched CMC/SWCNTfilms in which SWCNTs are partially aligned in the stretchdirection �not shown�. The whole PL spectrum of the unillu-minated samples showed an alignment ratio �5, in accor-dance with our previous results6,7 obtained for HiPCOSWCNTs. Remarkably, a similar alignment ratio �5 wasobserved for the UV-induced 1150 and 1240 nm peaks aswell. This result means that both the UV-induced and theoriginal electronic states have a similar anisotropic nature,implying electronic delocalization along the tube axis.The absorption spectra from the CMC/SWCNT sampleused for Figs. 1 and 2, before and after UV illumination, arepresented in Fig. 3. Right after D2 lamp illumination for15–180 min, the S11 peaks, appearing in the range900–1400 nm, but not the S22 and S33 peaks, are reduced byhalf. However, contrary to the irreversible PL changes occur-ring after �2 h illumination, the absorption changes alwaysdid recover within a few hours.In contrast to PL, PLE, and absorption results, no detect-able UV-induced changes were observed in Raman spectra�not shown�, excited at 632.8 nm, in the breathing, G or Dmodes.In the previous paragraphs, the UV illumination effectswere presented for CMC/SWCNT films. However, similaralterations could also be produced when CoMoCAT �orHiPCO� SWCNTs were dispersed with other agents�hydroxyethylcellulose,6 SDBS, sodium dodecylsulfate, etc.�.On the contrary, no PL changes could be detected when thedispersant was removed before UV illumination, showingthat the dispersant plays an essential role in this phenom-enon. It is worthwhile to note that the response time � andthe position of the UV-induced PL peaks varied significantlydepending on the dispersant �by a few tens of nanometer�.To reveal which spectral component of the UV light in-duces the PL changes, the D2 lamp was filtered with long-pass filters, whose cut-off wavelength was progressively de-creased. The thus obtained “PL quenching spectra” arepresented by squares and circles in Fig. 3 for two represen-tative samples: a CMC/SWCNT film and an SDBS/SWCNTsolution. These spectra exhibit abrupt thresholds at �250and 300 nm, which match the absorption thresholds of CMCand SDBS, respectively �lines in Fig. 3�. Here the PLquenching is defined as I�0� / I���−1, where I�t� is time-dependent PL intensity, which corresponds to the “revers-ible” part of the PL change.The data summarized above allow us to build a phenom-enological description of the UV-induced changes as follows.�1� The alterations in PL from SWCNTs are triggered by theUV excitation of the dispersant. This conclusion origi-FIG. 2. �Color online� PLE maps from a CMC/CoMoCAT film before �a�and after �b� UV illumination.FIG. 3. �Color online� Absorption spectra of a CMC/CoMoCAT film before�solid triangles� and after �open triangles� UV illumination. No extraSWCNT-related peaks could be detected in the range 1400–3200 nm. Solidand dashed lines present absorption spectra from 10 g/ l D2O solutions ofpure CMC and SDBS, respectively. Solid squares and open circles show PLquenching spectra in a CMC/CoMoCAT film and a SDBS/CoMoCAT solu-tion, respectively. The inset presents the D2 lamp spectrum.173108-2 Iakoubovskii et al. Appl. Phys. Lett. 89, 173108 �2006�Downloaded 24 Oct 2006 to 150.29.237.184. Redistribution subject to AIP license or copyright, see http://apl.aip.org/apl/copyright.jspnates from the close match between the excitation spec-tra of the PL quenching and the absorption spectra of thedispersant �Fig. 3�, and from the lack of PL changes inthe absence of dispersant. This result unambiguouslymanifests electronic interaction between the SWCNTsand UV-excited dispersants.�2� The time evolution of the UV-induced changes occurs intwo steps: �i� PL quenching �t��� and �ii� modificationof PL spectra �t���. The first step is fully reversible.The concomitant reduction in S11 absorption suggestsUV-induced electron injection into the conduction bands�or hole injection into the valence bands� of SWCNTs.We assume that these doping-induced, “metallic” statesinduce PL quenching by providing additional excitationrelaxation channels. Once UV illumination is termi-nated, reverse charge transfer occurs, fully restoring theoriginal absorption and PL. The larger reduction of PLthan of absorption signals could be explained by thestronger dependence of the PL on the relaxation process.�3� When UV illumination time exceeds �, the irreversible,second step comes into effect. We suggest that this pro-cess originates from the UV-induced, doped states, even-tually creating hitherto unknown luminescent defects.The PL peaks originating from these defects are not ob-servable immediately after the termination of this pro-longed UV illumination, possibly because PL quenchingby the doping-induced metallic states remains operative.As these metallic states disappear, the UV-induced PLpeaks emerge gradually �Fig. 1�.�4� The redshift of the PL peaks implies that these UV-induced defects have luminescent states in the forbiddengap of SWCNTs. The aforementioned polarization de-pendence of PL reveals strong anisotropy of the defectsalong the film stretching direction, indicating 1D delo-calized, rather than dotlike nature of these states. Notethat these UV-induced features appear broader than theparent peaks �see Fig. 2�, possibly due to the defect-related disorder, in agreement with the proposed thedefect-related origin.�5� The fact that no UV-induced features could be detectedin absorption �Fig. 3�, PLE �Fig. 2�, and Raman spectra�not shown� suggests that the SWCNT’s structure re-mains mostly intact and that the density of the createddefects is low. Nonetheless, the UV-induced PL peaksdominate over the original ones, probably because exci-tons migrating along 1D SWCNT can be very efficientlytrapped by these sparse defects.The present results have established the creation and ex-istence of midgap luminescent states in SWCNTs. While aUV-induced change in PL from 0.4 nm thick SWCNTsgrown in zeolite has been reported before,10 the range of UVintensity used ��100 W/cm2�, the way PL changes, and theunderlying mechanism seem completely different from thoseobserved here. It is rather surprising that relatively weak UVillumination ��1 mW/cm2� can induce such drastic PLchanges, provided that SWCNTs coexist with a dispersant.This observation suggests necessity for precaution againstSWCNT degradation when handling dispersed SWCNTsamples. On the positive side, however, these results couldbe of technological importance as they provide unprec-edented means to tune the emission wavelength of SWCNTs,as we indeed found that different dispersants resulted in dif-ferently shifted PL peaks.The authors are grateful to T. Okazaki for the use of theFluorolog spectrofluorometer. One of the authors �Y.K.�thanks the financial support from the JSPS Postdoctoral Fel-lowship for Foreign Researchers.1M. J. O’Connell, S. M. Bachilo, C. B. Huffman, V. C. Moore, M. S.Strano, E. H. Haroz, K. L. Rialon, P. J. Boul, W. H. Noon, C. Kittrell, J.Ma, R. H. Hauge, R. B. Weisman, and R. E. Smalley, Science 297, 593�2002�.2F. Wang, G. Dukovic, L. E. Brus, and T. F. Heinz, Science 308, 838�2005�.3G. Dukovic, F. Wang, D. Song, M. Y. Sfeir, T. F. Heinz, and L. E. Brus,Nano Lett. 5, 2314 �2005�.4J. Lefebre, Y. Homma, and P. Finnie, Phys. Rev. Lett. 90, 217401 �2003�.5J. Guo, C. Yang, Z. M. Li, M. Bai, H. J. Liu, G. D. Li, E. G. Wang, C. T.Chan, Z. K. Tang, W. K. Ge, and X. Xiao, Phys. Rev. Lett. 93, 017402�2004�.6N. Minami, Y. Kim, K. Miyashita, B. Nalini, and S. Kazaoui, Appl. Phys.Lett. 88, 93123 �2006�.7Y. Kim, N. Minami, and S. Kazaoui, Appl. Phys. Lett. 86, 73103 �2005�.8S. G. Chou, F. Plentz, J. Jiang, R. Saito, D. Nezich, H. B. Ribeiro, A.Jorio, M. A. Pimenta, G. G. Samsonidze, A. P. Santos, M. Zheng, G. B.Onoa, E. D. Semke, G. Dresselhaus, and M. S. Dresselhaus, Phys. Rev.Lett. 94, 127402 �2005�.9S. M. Bachilo, M. S. Strano, C. Kittrell, R. H. Hauge, R. E. Smalley, andR. B. Weisman, Science 298, 2361 �2002�.10M. Bai, I. L. Li, Z. K. Tang, and X. Xiao, Appl. Phys. Lett. 86, 93108�2005�.173108-3 Iakoubovskii et al. Appl. Phys. Lett. 89, 173108 �2006�Downloaded 24 Oct 2006 to 150.29.237.184. Redistribution subject to AIP license or copyright, see http://apl.aip.org/apl/copyright.jsp